Dell Finally Goes for AMD 278
this great guy writes "You read it correctly. It had to happen one day.
According to Forbes
'Dell Inc has informed its Taiwan contract makers of plans to develop devices based on Advanced Micro Devices Inc's microprocessors, and these suppliers are awaiting orders for global shipment, the Economic Daily News reported, citing industry sources.'"
intel... (Score:5, Interesting)
It's surprising, but really should give hope to the Linux crew. This is a testament that in our society the underdog can come up and take a bite out of the front runner.... regardless of their fortunes. look out ms.
Debateable point? (Score:3, Interesting)
To me, this makes little sense. I think we can all agree that the Pentium-M is superior for laptops, so there is little point in Dell producing AMD based laptops, especially given the AMD ones will be more expensive once subsidies from Intel are accounted for.
Desktops are a similar story - there AMD has the superior processor, but it's still going to wind up more expensive thanks to Intel.
Servers are the only market where this seems to be a good move, as it will allow Dell to flog dual-core server rigs which Intel are currently unable to provide.
Unless I'm mising something of course...
Re:Nothing but good... (Score:5, Interesting)
This is exactly what the law suit was intended to do. I couldn't disagree more with your assessment concerning the validity of the suit.
It didn't really have to happen (Score:5, Interesting)
I'd assume that because of this fact, AMD will most likely only appear in servers where Intel can't deliver, because if they go down to the desktop level, and there's no product differentiation, Dell will most likely find themselves in the same position as HP, who I don't buy from simply because I don't have the time to fathom the differences between 47 desktop models, 37 laptops, 53 laser printers and 73 inkjets with varying specs.
Dell won out for me because they kept it simple and focused, I hope they still do that, rather than end up fighting AMD vs. Intel between their own products as others do.
I wonder if the sales losses are mounting? (Score:5, Interesting)
Let's assume for an instant that the story is correct.
It could be that Dell have started noticing that they're losing out on contracts because of the heat that their Intel based servers are kicking out. I know of several contracts (even in my limited circle of knowledge) where the Dell submision appeared to give the most bang-per-buck, but didn't get the contract when the cost of the extra air-con was included in the total cost of ownership.
Air-con is a non-negligible expense in any half-way serious machine room.
I wonder if this does have some degree of merit... (Score:2, Interesting)
Re:price/performance (Score:2, Interesting)
AMD History [amd.com]
Re:Loyalty (Score:5, Interesting)
Desktop sales are dropping. Laptop sales are growing. The two lines on the graph are crossing right about now. Next year, laptop sales are projected to outnumber desktop sales, and keep growing. I think Intel are exactly right to bet on the laptop market. AMD are mainly targetting the supercomputer, server, and workstation markets. These are low-volume, high-margin areas, and are ideally suited to a company with a good R&D team (lots of ex-Alpha people) and a lower volume production capacity.
Re:price/performance (Score:3, Interesting)
Re:Loyalty (Score:2, Interesting)
The mistakes of offshoring coming home to roost (Score:1, Interesting)
Their current problem is due to bad engineering. They just don't have the engineering talent to pull themselves out of the dumper anymore. And the reason why they don't have the talent is because they have aggressively offshored their jobs, at the expense of bringing in new talent.
Had they actually been hiring the brightest new kids out of college and training them under the guidance of experienced engineers, Intel wouldn't be in this mess. But no, the PHB's have fallen for the H1/L1/offshore hype in a big way. I'm sorry, I've worked a lot with the people from India, and 99% of them just don't cut it (the other 1% is impressive, but they always end up with better opportunities elsewhere).
Now it's time to pay the piper.
Intel deserves its current course straight into the dumper. And they no longer have the technical talent to do anything great anymore; it's questionable if they have the ability to reverse the course they're on.
It's absolutely delightful to see the mistakes of offshoring coming back to bite Intel in a very big way.
Re:Nothing but good... (Score:4, Interesting)
Here's the story as I rememnber it, and I'm sure there are others with more inside information and better memories out there who can complete the story:
Intel equated processing power with core speed. This began a kind of arms race between AMD and Intel in clock speed - but the AMD message was diluted because they focussed on better chips, not just better scores. Several years ago, however, they decided to play at Intel's own game for a time and there was this huge rush to get to the 1 GHZ mark. It was in this frenzy of clock speed that Intel switched from the efficient and well-designed p3 (which they later returned to for their excellent Pentium M chip) to their wimpy but fast p4. They were able to push the p4 speeds higher than the p3 speeds and thus continue to win the "clock speed = power" battle with the public.
It was on the foundation of this general misconception that Intel's brand was built (of course the actual marketing - as in commercials on TV - had practically no technical content whatsoever, so I'm talking about marketting in terms of their development strategy to win over the computer users who were looking for some simple number they could relate to value).
Now, of course, with the rising supremacy of mobile computing (where the PM shines), the advent of 64-bit processors (AMD was the leader) and dual core processors (again, kudos to AMD) they have FINALLY instituted a marketting shift. This was apparent months ago when they unveiled their new road map and everything was about cycles/watt. They've got a new, slighty more complex and slightly better marketting gimmick - but it's the same old idea. The public doesn't want to read dozens of Anandtech articles detailing each new core for each new processor - they just want a guideline that makes sense. For a while it was clock speed, now it's clock speed per watt.
You have to give them credit for clever marketting. It was because of ploys like this that their brand name became the defacto standard in CPUs, but it doesn't alter the fact that it's just a marketing ploy and that Intel has used their ability to misrepresent their chips to the detriment of actual chip design. It it weren't for AMD - where would we be now?
Although in the final analysis - the real savior isn't AMD, it's competition.
-stormin
Re:Loyalty (Score:3, Interesting)
Re:intel... (Score:1, Interesting)
I work for Intel, and I just want to correct some misconceptions.
Yes, the P4 is now slower than the Opertons. But not by a huge amount.
The real problem is that we underestimated the power scaling of the architecture. The P4 was designed to scale up to 6-7 GHz (Hence the long pipeline). In fact, at this time right now, we had plans for a 6GHz P4 which would surely top anything AMD has right now. I mean, if the pipeline is 2x longer but you can run the chip 3x faster then you come out ahead (these are hypothetical numbers but you get the idea behind it).
The reason why you don't have 6GHz processors today is that the pre-silicon power models grossly underpredicted reality. We shot for the moon and fell short. That's the price of risk-taking and doing something innovative. AMD played it somewhat "safe" and is reaping the benefits.
HT is good, not bad - it allows you to efficiently put idle resources to use. Problems can be corrected with more intelligent algorithms on how to dynamically allocate shared resources amongst threads. We were late on dual core and general system architecture goodness because lots of our engineers were busy pissing away their time on Itanium.
Intel has a near monopoly on microprocessor engineers and management can only be clueless for so long before the "market" eliminates them. Intel also has about a 9 month process step lead over everybody.
We'll see who comes out ahead in the post-P4 battle... My money is on Intel.